Análisis de modelos SOM - Frecuencia datos de entrada: mes

Modelo

• ID: 257
• Descripción:
• Frecuencia: mes
• Variables: fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud
• Dimensiones del mapa: 5,5
• Iteraciones: 1000
• Parámetros adicionales:

source("../../lib/som-utils.R")

Attaching package: 'dplyr'

The following objects are masked from 'package:stats':

    filter, lag

The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union

source("../../lib/maps-utils.R")

Linking to GEOS 3.8.0, GDAL 3.0.4, PROJ 6.3.1

Carga del modelo desde disco

mpr.set_base_path_analysis()
model <- mpr.load_model("som-257.rds.xz")
summary(model)

SOM of size 5x5 with a hexagonal topology and a bubble neighbourhood function.
The number of data layers is 1.
Distance measure(s) used: sumofsquares.
Training data included: 94881 objects.
Mean distance to the closest unit in the map: 1.827.

plot(model, type="changes")

Carga del dataset de entrada

df <- mpr.load_data("datos_mes.csv.xz")

df

summary(df)

 id_estacion           fecha             fecha_cnt           tmax      
 Length:94881       Length:94881       Min.   : 1.000   Min.   :-53.0  
 Class :character   Class :character   1st Qu.: 4.000   1st Qu.:148.0  
 Mode  :character   Mode  :character   Median : 6.000   Median :198.0  
                                       Mean   : 6.497   Mean   :200.2  
                                       3rd Qu.: 9.000   3rd Qu.:255.0  
                                       Max.   :12.000   Max.   :403.0  
      tmin             precip           nevada           prof_nieve      
 Min.   :-121.00   Min.   :  0.00   Min.   :0.000000   Min.   :   0.000  
 1st Qu.:  53.00   1st Qu.:  3.00   1st Qu.:0.000000   1st Qu.:   0.000  
 Median :  98.00   Median : 10.00   Median :0.000000   Median :   0.000  
 Mean   :  98.86   Mean   : 16.25   Mean   :0.000295   Mean   :   0.467  
 3rd Qu.: 148.00   3rd Qu.: 22.00   3rd Qu.:0.000000   3rd Qu.:   0.000  
 Max.   : 254.00   Max.   :422.00   Max.   :6.000000   Max.   :1834.000  
    longitud        latitud            altitud      
 Min.   :27.82   Min.   :-17.8889   Min.   :   1.0  
 1st Qu.:38.28   1st Qu.: -5.6417   1st Qu.:  42.0  
 Median :40.82   Median : -3.4500   Median : 247.0  
 Mean   :39.66   Mean   : -3.4350   Mean   : 418.5  
 3rd Qu.:42.08   3rd Qu.:  0.4914   3rd Qu.: 656.0  
 Max.   :43.57   Max.   :  4.2156   Max.   :2535.0  

Carga de los mapas

world <- ne_countries(scale = "medium", returnclass = "sf")
spain <- subset(world, admin == "Spain")

Mapa de densidad

plot(model, type="count", shape = "straight", palette.name = mpr.degrade.bleu)

Número de elementos en cada celda:

nb <- table(model$unit.classif)
print(nb)

   1    2    3    4    5    6    7    8    9   10   11   12   13   14   15   16 
7677 5060 4615 5850    9 7573 3172 2435 4601 3107 5251 3885 3231 1855  746 5680 
  17   18   19   20   21   22   23   24   25 
2852 4940 5339 2548 1189 4399 2852 5985   30 

Comprobación de nodos vacíos:

dim_model <- 5*5;
len_nb = length(nb);
empty_nodes <- dim_model != len_nb;
if (empty_nodes) {
  print(paste("[Warning] Existen nodos vacíos: ", len_nb, "/", dim_model))
}

Mapa de distancia entre vecinos

plot(model, type="dist.neighbours", shape = "straight")

Influencia de las variables

model_colnames = c("fecha_cnt", "tmax", "tmin", "precip", "nevada", "prof_nieve", "longitud", "latitud", "altitud")
model_ncol = length(model_colnames)

Mapa de variables.

plot(model, shape = "straight")

Mapa de calor por variable

par(mfrow=c(3,4))
for (j in 1:model_ncol) {
  plot(model, type="property", property=getCodes(model,1)[,j],
    palette.name=mpr.coolBlueHotRed,
    main=model_colnames[j],
    cex=0.5, shape = "straight")
}

Correlación para cada columna del vector de nodos

if (!empty_nodes) {
  cor <- apply(getCodes(model,1), 2, mpr.weighted.correlation, w=nb, som=model)
  print(cor)
}

      fecha_cnt       tmax       tmin      precip      nevada prof_nieve
[1,] -0.1326450 -0.7764810 -0.7881419  0.50837115  0.01949938 0.04073106
[2,] -0.7699525 -0.3342997 -0.2678378 -0.01103149 -0.01261588 0.02232395
       longitud    latitud     altitud
[1,]  0.5567297  0.2003750  0.15767241
[2,] -0.3705334 -0.2755058 -0.08013338

Representación de cada variable en un mapa de factores:

if (!empty_nodes) {
  par(mfrow=c(1,1))
  plot(cor[1,], cor[2,], xlim=c(-1,1), ylim=c(-1,1), type="n")
  lines(c(-1,1),c(0,0))
  lines(c(0,0),c(-1,1))
  text(cor[1,], cor[2,], labels=model_colnames, cex=0.75)
  symbols(0,0,circles=1,inches=F,add=T)
}

Importancia de cada variable - varianza ponderada por el tamaño de la celda:

if (!empty_nodes) {
  sigma2 <- sqrt(apply(getCodes(model,1),2,function(x,effectif)
     {m<-sum(effectif*(x-weighted.mean(x,effectif))^2)/(sum(effectif)-1)},
     effectif=nb))
  print(sort(sigma2,decreasing=T))
}

  longitud    altitud       tmax    latitud       tmin  fecha_cnt     nevada 
 0.9413777  0.9196386  0.9146835  0.9135709  0.9104839  0.8990644  0.8769013 
    precip prof_nieve 
 0.8325967  0.8284927 

Clustering

if (!empty_nodes) {
  hac <- mpr.hac(model, nb)
}

Visualización de 3 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=3)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=3)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax            tmin             precip      
 Min.   : 1.000   Min.   :-53.0   Min.   :-121.00   Min.   :  0.00  
 1st Qu.: 4.000   1st Qu.:149.0   1st Qu.:  53.00   1st Qu.:  3.00  
 Median : 6.000   Median :198.0   Median :  98.00   Median : 10.00  
 Mean   : 6.498   Mean   :200.3   Mean   :  98.91   Mean   : 16.24  
 3rd Qu.: 9.000   3rd Qu.:255.0   3rd Qu.: 148.00   3rd Qu.: 22.00  
 Max.   :12.000   Max.   :403.0   Max.   : 254.00   Max.   :422.00  
     nevada           prof_nieve          longitud        latitud        
 Min.   :0.00e+00   Min.   :  0.0000   Min.   :27.82   Min.   :-17.8889  
 1st Qu.:0.00e+00   1st Qu.:  0.0000   1st Qu.:38.28   1st Qu.: -5.6417  
 Median :0.00e+00   Median :  0.0000   Median :40.82   Median : -3.4500  
 Mean   :3.16e-05   Mean   :  0.2499   Mean   :39.66   Mean   : -3.4350  
 3rd Qu.:0.00e+00   3rd Qu.:  0.0000   3rd Qu.:42.08   3rd Qu.:  0.4914  
 Max.   :1.00e+00   Max.   :343.0000   Max.   :43.57   Max.   :  4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  42.0  
 Median : 247.0  
 Mean   : 418.1  
 3rd Qu.: 656.0  
 Max.   :2535.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt         tmax             tmin            precip           nevada 
 Min.   : 1.0   Min.   :-12.00   Min.   :-72.00   Min.   : 16.00   Min.   :0  
 1st Qu.: 2.0   1st Qu.: 10.25   1st Qu.:-43.75   1st Qu.: 32.25   1st Qu.:0  
 Median : 2.5   Median : 24.00   Median :-29.50   Median : 50.00   Median :0  
 Mean   : 2.7   Mean   : 24.30   Mean   :-31.17   Mean   : 56.10   Mean   :0  
 3rd Qu.: 3.0   3rd Qu.: 35.75   3rd Qu.:-23.25   3rd Qu.: 71.50   3rd Qu.:0  
 Max.   :12.0   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve        longitud        latitud         altitud    
 Min.   : 357.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.: 481.5   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median : 599.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   : 685.7   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.: 765.2   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

Visualización de 4 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=4)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=4)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax            tmin             precip      
 Min.   : 1.000   Min.   :-53.0   Min.   :-121.00   Min.   :  0.00  
 1st Qu.: 4.000   1st Qu.:146.0   1st Qu.:  50.00   1st Qu.:  3.00  
 Median : 6.000   Median :194.0   Median :  93.00   Median : 11.00  
 Mean   : 6.498   Mean   :198.9   Mean   :  94.87   Mean   : 16.96  
 3rd Qu.: 9.000   3rd Qu.:255.0   3rd Qu.: 141.00   3rd Qu.: 23.00  
 Max.   :12.000   Max.   :403.0   Max.   : 254.00   Max.   :422.00  
     nevada           prof_nieve         longitud        latitud        
 Min.   :0.00e+00   Min.   :  0.000   Min.   :27.82   Min.   :-17.8889  
 1st Qu.:0.00e+00   1st Qu.:  0.000   1st Qu.:38.99   1st Qu.: -4.8500  
 Median :0.00e+00   Median :  0.000   Median :40.96   Median : -2.4831  
 Mean   :3.41e-05   Mean   :  0.267   Mean   :40.55   Mean   : -2.4501  
 3rd Qu.:0.00e+00   3rd Qu.:  0.000   3rd Qu.:42.23   3rd Qu.:  0.4942  
 Max.   :1.00e+00   Max.   :343.000   Max.   :43.57   Max.   :  4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  44.0  
 Median : 251.0  
 Mean   : 410.2  
 3rd Qu.: 667.0  
 Max.   :2535.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt           tmax            tmin           precip            nevada 
 Min.   : 1.000   Min.   : 13.0   Min.   :-33.0   Min.   :  0.000   Min.   :0  
 1st Qu.: 4.000   1st Qu.:201.0   1st Qu.:132.0   1st Qu.:  0.000   1st Qu.:0  
 Median : 7.000   Median :225.0   Median :160.0   Median :  1.000   Median :0  
 Mean   : 6.503   Mean   :217.4   Mean   :150.6   Mean   :  6.924   Mean   :0  
 3rd Qu.: 9.000   3rd Qu.:253.0   3rd Qu.:190.0   3rd Qu.:  8.000   3rd Qu.:0  
 Max.   :12.000   Max.   :356.0   Max.   :244.0   Max.   :114.000   Max.   :0  
   prof_nieve          longitud        latitud          altitud      
 Min.   : 0.00000   Min.   :27.82   Min.   :-17.89   Min.   :  14.0  
 1st Qu.: 0.00000   1st Qu.:28.31   1st Qu.:-16.50   1st Qu.:  25.0  
 Median : 0.00000   Median :28.44   Median :-16.33   Median :  35.0  
 Mean   : 0.03086   Mean   :28.36   Mean   :-16.05   Mean   : 518.7  
 3rd Qu.: 0.00000   3rd Qu.:28.48   3rd Qu.:-15.39   3rd Qu.: 632.0  
 Max.   :46.00000   Max.   :28.95   Max.   :-13.60   Max.   :2371.0  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt         tmax             tmin            precip           nevada 
 Min.   : 1.0   Min.   :-12.00   Min.   :-72.00   Min.   : 16.00   Min.   :0  
 1st Qu.: 2.0   1st Qu.: 10.25   1st Qu.:-43.75   1st Qu.: 32.25   1st Qu.:0  
 Median : 2.5   Median : 24.00   Median :-29.50   Median : 50.00   Median :0  
 Mean   : 2.7   Mean   : 24.30   Mean   :-31.17   Mean   : 56.10   Mean   :0  
 3rd Qu.: 3.0   3rd Qu.: 35.75   3rd Qu.:-23.25   3rd Qu.: 71.50   3rd Qu.:0  
 Max.   :12.0   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve        longitud        latitud         altitud    
 Min.   : 357.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.: 481.5   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median : 599.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   : 685.7   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.: 765.2   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

Visualización de 5 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=5)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=5)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)
  df.cluster05 <- subset(df, cluster==5)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster05 <- select(df.cluster05, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax          tmin            precip     
 Min.   : 1.000   Min.   :  8   Min.   :-110.0   Min.   : 0.00  
 1st Qu.: 7.000   1st Qu.:194   1st Qu.:  92.0   1st Qu.: 2.00  
 Median : 8.000   Median :243   Median : 132.0   Median : 8.00  
 Mean   : 8.493   Mean   :235   Mean   : 124.6   Mean   :11.38  
 3rd Qu.:10.000   3rd Qu.:285   3rd Qu.: 165.0   3rd Qu.:17.00  
 Max.   :12.000   Max.   :403   Max.   : 254.0   Max.   :95.00  
     nevada           prof_nieve         longitud        latitud       
 Min.   :0.00e+00   Min.   : 0.0000   Min.   :35.28   Min.   :-8.6494  
 1st Qu.:0.00e+00   1st Qu.: 0.0000   1st Qu.:38.88   1st Qu.:-4.6992  
 Median :0.00e+00   Median : 0.0000   Median :40.82   Median :-2.1381  
 Mean   :3.96e-05   Mean   : 0.0172   Mean   :40.33   Mean   :-2.1936  
 3rd Qu.:0.00e+00   3rd Qu.: 0.0000   3rd Qu.:41.84   3rd Qu.: 0.5706  
 Max.   :1.00e+00   Max.   :59.0000   Max.   :43.57   Max.   : 4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  43.0  
 Median : 192.0  
 Mean   : 353.5  
 3rd Qu.: 628.0  
 Max.   :1971.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt           tmax            tmin             precip     
 Min.   : 1.000   Min.   :-53.0   Min.   :-121.00   Min.   :  0.0  
 1st Qu.: 2.000   1st Qu.:123.0   1st Qu.:  28.00   1st Qu.:  7.0  
 Median : 3.000   Median :155.0   Median :  59.00   Median : 17.0  
 Mean   : 3.804   Mean   :150.3   Mean   :  54.74   Mean   : 24.5  
 3rd Qu.: 4.000   3rd Qu.:181.0   3rd Qu.:  83.00   3rd Qu.: 34.0  
 Max.   :12.000   Max.   :350.0   Max.   : 223.00   Max.   :422.0  
     nevada           prof_nieve          longitud        latitud        
 Min.   :0.00e+00   Min.   :  0.0000   Min.   :27.82   Min.   :-17.8889  
 1st Qu.:0.00e+00   1st Qu.:  0.0000   1st Qu.:39.49   1st Qu.: -5.5975  
 Median :0.00e+00   Median :  0.0000   Median :41.38   Median : -3.4503  
 Mean   :2.67e-05   Mean   :  0.6043   Mean   :40.83   Mean   : -2.7962  
 3rd Qu.:0.00e+00   3rd Qu.:  0.0000   3rd Qu.:42.47   3rd Qu.:  0.4731  
 Max.   :1.00e+00   Max.   :343.0000   Max.   :43.57   Max.   :  4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  58.0  
 Median : 336.0  
 Mean   : 486.8  
 3rd Qu.: 704.0  
 Max.   :2535.0  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt           tmax            tmin           precip            nevada 
 Min.   : 1.000   Min.   : 13.0   Min.   :-33.0   Min.   :  0.000   Min.   :0  
 1st Qu.: 4.000   1st Qu.:201.0   1st Qu.:132.0   1st Qu.:  0.000   1st Qu.:0  
 Median : 7.000   Median :225.0   Median :160.0   Median :  1.000   Median :0  
 Mean   : 6.503   Mean   :217.4   Mean   :150.6   Mean   :  6.924   Mean   :0  
 3rd Qu.: 9.000   3rd Qu.:253.0   3rd Qu.:190.0   3rd Qu.:  8.000   3rd Qu.:0  
 Max.   :12.000   Max.   :356.0   Max.   :244.0   Max.   :114.000   Max.   :0  
   prof_nieve          longitud        latitud          altitud      
 Min.   : 0.00000   Min.   :27.82   Min.   :-17.89   Min.   :  14.0  
 1st Qu.: 0.00000   1st Qu.:28.31   1st Qu.:-16.50   1st Qu.:  25.0  
 Median : 0.00000   Median :28.44   Median :-16.33   Median :  35.0  
 Mean   : 0.03086   Mean   :28.36   Mean   :-16.05   Mean   : 518.7  
 3rd Qu.: 0.00000   3rd Qu.:28.48   3rd Qu.:-15.39   3rd Qu.: 632.0  
 Max.   :46.00000   Max.   :28.95   Max.   :-13.60   Max.   :2371.0  

if (!empty_nodes) summary(df.cluster05)

   fecha_cnt         tmax             tmin            precip           nevada 
 Min.   : 1.0   Min.   :-12.00   Min.   :-72.00   Min.   : 16.00   Min.   :0  
 1st Qu.: 2.0   1st Qu.: 10.25   1st Qu.:-43.75   1st Qu.: 32.25   1st Qu.:0  
 Median : 2.5   Median : 24.00   Median :-29.50   Median : 50.00   Median :0  
 Mean   : 2.7   Mean   : 24.30   Mean   :-31.17   Mean   : 56.10   Mean   :0  
 3rd Qu.: 3.0   3rd Qu.: 35.75   3rd Qu.:-23.25   3rd Qu.: 71.50   3rd Qu.:0  
 Max.   :12.0   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve        longitud        latitud         altitud    
 Min.   : 357.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.: 481.5   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median : 599.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   : 685.7   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.: 765.2   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1], dim(df.cluster05)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04", "cluster05"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.hist(df.cluster05)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster05)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
  df.cluster05.grouped <- mpr.group_by_geo(df.cluster05)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster05.grouped)

Visualización de 6 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=6)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=6)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)
  df.cluster05 <- subset(df, cluster==5)
  df.cluster06 <- subset(df, cluster==6)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster05 <- select(df.cluster05, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster06 <- select(df.cluster06, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax          tmin            precip     
 Min.   : 1.000   Min.   :  8   Min.   :-110.0   Min.   : 0.00  
 1st Qu.: 7.000   1st Qu.:194   1st Qu.:  92.0   1st Qu.: 2.00  
 Median : 8.000   Median :243   Median : 132.0   Median : 8.00  
 Mean   : 8.493   Mean   :235   Mean   : 124.6   Mean   :11.38  
 3rd Qu.:10.000   3rd Qu.:285   3rd Qu.: 165.0   3rd Qu.:17.00  
 Max.   :12.000   Max.   :403   Max.   : 254.0   Max.   :95.00  
     nevada           prof_nieve         longitud        latitud       
 Min.   :0.00e+00   Min.   : 0.0000   Min.   :35.28   Min.   :-8.6494  
 1st Qu.:0.00e+00   1st Qu.: 0.0000   1st Qu.:38.88   1st Qu.:-4.6992  
 Median :0.00e+00   Median : 0.0000   Median :40.82   Median :-2.1381  
 Mean   :3.96e-05   Mean   : 0.0172   Mean   :40.33   Mean   :-2.1936  
 3rd Qu.:0.00e+00   3rd Qu.: 0.0000   3rd Qu.:41.84   3rd Qu.: 0.5706  
 Max.   :1.00e+00   Max.   :59.0000   Max.   :43.57   Max.   : 4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  43.0  
 Median : 192.0  
 Mean   : 353.5  
 3rd Qu.: 628.0  
 Max.   :1971.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt           tmax          tmin             precip     
 Min.   : 1.000   Min.   :-53   Min.   :-121.00   Min.   :  0.0  
 1st Qu.: 2.000   1st Qu.:122   1st Qu.:  23.00   1st Qu.:  6.0  
 Median : 3.000   Median :156   Median :  54.00   Median : 13.0  
 Mean   : 3.081   Mean   :150   Mean   :  50.27   Mean   : 16.7  
 3rd Qu.: 4.000   3rd Qu.:182   3rd Qu.:  80.00   3rd Qu.: 24.0  
 Max.   :12.000   Max.   :281   Max.   : 168.00   Max.   :126.0  
     nevada           prof_nieve          longitud        latitud       
 Min.   :0.00e+00   Min.   :  0.0000   Min.   :35.28   Min.   :-8.6494  
 1st Qu.:0.00e+00   1st Qu.:  0.0000   1st Qu.:39.47   1st Qu.:-4.8458  
 Median :0.00e+00   Median :  0.0000   Median :41.06   Median :-2.7331  
 Mean   :3.15e-05   Mean   :  0.6906   Mean   :40.63   Mean   :-2.3933  
 3rd Qu.:0.00e+00   3rd Qu.:  0.0000   3rd Qu.:42.26   3rd Qu.: 0.5933  
 Max.   :1.00e+00   Max.   :343.0000   Max.   :43.57   Max.   : 4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  64.0  
 Median : 427.0  
 Mean   : 536.2  
 3rd Qu.: 779.0  
 Max.   :2535.0  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt           tmax            tmin            precip           nevada 
 Min.   : 1.000   Min.   : -8.0   Min.   :-58.00   Min.   : 20.00   Min.   :0  
 1st Qu.: 3.000   1st Qu.:127.0   1st Qu.: 58.00   1st Qu.: 49.00   1st Qu.:0  
 Median :10.000   Median :149.0   Median : 77.00   Median : 60.00   Median :0  
 Mean   : 7.823   Mean   :152.1   Mean   : 79.58   Mean   : 67.84   Mean   :0  
 3rd Qu.:11.000   3rd Qu.:175.0   3rd Qu.:100.00   3rd Qu.: 79.00   3rd Qu.:0  
 Max.   :12.000   Max.   :350.0   Max.   :223.00   Max.   :422.00   Max.   :0  
   prof_nieve         longitud        latitud           altitud      
 Min.   : 0.0000   Min.   :27.82   Min.   :-17.889   Min.   :   1.0  
 1st Qu.: 0.0000   1st Qu.:41.98   1st Qu.: -8.372   1st Qu.:  42.0  
 Median : 0.0000   Median :42.89   Median : -5.642   Median : 108.0  
 Mean   : 0.1246   Mean   :41.95   Mean   : -5.035   Mean   : 211.9  
 3rd Qu.: 0.0000   3rd Qu.:43.36   3rd Qu.: -2.906   3rd Qu.: 336.0  
 Max.   :81.0000   Max.   :43.57   Max.   :  4.216   Max.   :2400.0  

if (!empty_nodes) summary(df.cluster05)

   fecha_cnt           tmax            tmin           precip            nevada 
 Min.   : 1.000   Min.   : 13.0   Min.   :-33.0   Min.   :  0.000   Min.   :0  
 1st Qu.: 4.000   1st Qu.:201.0   1st Qu.:132.0   1st Qu.:  0.000   1st Qu.:0  
 Median : 7.000   Median :225.0   Median :160.0   Median :  1.000   Median :0  
 Mean   : 6.503   Mean   :217.4   Mean   :150.6   Mean   :  6.924   Mean   :0  
 3rd Qu.: 9.000   3rd Qu.:253.0   3rd Qu.:190.0   3rd Qu.:  8.000   3rd Qu.:0  
 Max.   :12.000   Max.   :356.0   Max.   :244.0   Max.   :114.000   Max.   :0  
   prof_nieve          longitud        latitud          altitud      
 Min.   : 0.00000   Min.   :27.82   Min.   :-17.89   Min.   :  14.0  
 1st Qu.: 0.00000   1st Qu.:28.31   1st Qu.:-16.50   1st Qu.:  25.0  
 Median : 0.00000   Median :28.44   Median :-16.33   Median :  35.0  
 Mean   : 0.03086   Mean   :28.36   Mean   :-16.05   Mean   : 518.7  
 3rd Qu.: 0.00000   3rd Qu.:28.48   3rd Qu.:-15.39   3rd Qu.: 632.0  
 Max.   :46.00000   Max.   :28.95   Max.   :-13.60   Max.   :2371.0  

if (!empty_nodes) summary(df.cluster06)

   fecha_cnt         tmax             tmin            precip           nevada 
 Min.   : 1.0   Min.   :-12.00   Min.   :-72.00   Min.   : 16.00   Min.   :0  
 1st Qu.: 2.0   1st Qu.: 10.25   1st Qu.:-43.75   1st Qu.: 32.25   1st Qu.:0  
 Median : 2.5   Median : 24.00   Median :-29.50   Median : 50.00   Median :0  
 Mean   : 2.7   Mean   : 24.30   Mean   :-31.17   Mean   : 56.10   Mean   :0  
 3rd Qu.: 3.0   3rd Qu.: 35.75   3rd Qu.:-23.25   3rd Qu.: 71.50   3rd Qu.:0  
 Max.   :12.0   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve        longitud        latitud         altitud    
 Min.   : 357.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.: 481.5   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median : 599.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   : 685.7   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.: 765.2   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1], dim(df.cluster05)[1], dim(df.cluster06)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04", "cluster05", "cluster06"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.hist(df.cluster05)

if (!empty_nodes) mpr.hist(df.cluster06)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster05)

if (!empty_nodes) mpr.boxplot(df.cluster06)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
  df.cluster05.grouped <- mpr.group_by_geo(df.cluster05)
  df.cluster06.grouped <- mpr.group_by_geo(df.cluster06)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster05.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster06.grouped)

Visualización de 8 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=8)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=8)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)
  df.cluster05 <- subset(df, cluster==5)
  df.cluster06 <- subset(df, cluster==6)
  df.cluster07 <- subset(df, cluster==7)
  df.cluster08 <- subset(df, cluster==8)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster05 <- select(df.cluster05, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster06 <- select(df.cluster06, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster07 <- select(df.cluster07, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster08 <- select(df.cluster08, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax          tmin            precip     
 Min.   : 1.000   Min.   :  8   Min.   :-110.0   Min.   : 0.00  
 1st Qu.: 7.000   1st Qu.:194   1st Qu.:  92.0   1st Qu.: 2.00  
 Median : 8.000   Median :243   Median : 132.0   Median : 8.00  
 Mean   : 8.493   Mean   :235   Mean   : 124.6   Mean   :11.38  
 3rd Qu.:10.000   3rd Qu.:285   3rd Qu.: 165.0   3rd Qu.:17.00  
 Max.   :12.000   Max.   :403   Max.   : 254.0   Max.   :95.00  
     nevada           prof_nieve         longitud        latitud       
 Min.   :0.00e+00   Min.   : 0.0000   Min.   :35.28   Min.   :-8.6494  
 1st Qu.:0.00e+00   1st Qu.: 0.0000   1st Qu.:38.88   1st Qu.:-4.6992  
 Median :0.00e+00   Median : 0.0000   Median :40.82   Median :-2.1381  
 Mean   :3.96e-05   Mean   : 0.0172   Mean   :40.33   Mean   :-2.1936  
 3rd Qu.:0.00e+00   3rd Qu.: 0.0000   3rd Qu.:41.84   3rd Qu.: 0.5706  
 Max.   :1.00e+00   Max.   :59.0000   Max.   :43.57   Max.   : 4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  43.0  
 Median : 192.0  
 Mean   : 353.5  
 3rd Qu.: 628.0  
 Max.   :1971.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt           tmax             tmin             precip      
 Min.   : 1.000   Min.   :-53.00   Min.   :-121.00   Min.   :  0.00  
 1st Qu.: 3.000   1st Qu.: 27.00   1st Qu.: -35.00   1st Qu.: 20.00  
 Median : 6.000   Median : 66.00   Median :   1.00   Median : 32.00  
 Mean   : 6.425   Mean   : 72.75   Mean   :   5.09   Mean   : 36.21  
 3rd Qu.:10.000   3rd Qu.:117.00   3rd Qu.:  45.00   3rd Qu.: 48.00  
 Max.   :12.000   Max.   :218.00   Max.   : 128.00   Max.   :126.00  
     nevada    prof_nieve         longitud        latitud           altitud    
 Min.   :0   Min.   :  0.000   Min.   :40.78   Min.   :-4.0103   Min.   :1213  
 1st Qu.:0   1st Qu.:  0.000   1st Qu.:41.77   1st Qu.: 0.7292   1st Qu.:1894  
 Median :0   Median :  0.000   Median :42.47   Median : 0.9844   Median :2230  
 Mean   :0   Mean   :  7.832   Mean   :42.06   Mean   : 0.0693   Mean   :2160  
 3rd Qu.:0   3rd Qu.:  0.000   3rd Qu.:42.64   3rd Qu.: 1.5242   3rd Qu.:2400  
 Max.   :0   Max.   :343.000   Max.   :42.77   Max.   : 2.4378   Max.   :2535  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt           tmax            tmin            precip           nevada 
 Min.   : 1.000   Min.   : -8.0   Min.   :-58.00   Min.   : 20.00   Min.   :0  
 1st Qu.: 3.000   1st Qu.:127.0   1st Qu.: 58.00   1st Qu.: 49.00   1st Qu.:0  
 Median :10.000   Median :149.0   Median : 77.00   Median : 60.00   Median :0  
 Mean   : 7.823   Mean   :152.1   Mean   : 79.58   Mean   : 67.84   Mean   :0  
 3rd Qu.:11.000   3rd Qu.:175.0   3rd Qu.:100.00   3rd Qu.: 79.00   3rd Qu.:0  
 Max.   :12.000   Max.   :350.0   Max.   :223.00   Max.   :422.00   Max.   :0  
   prof_nieve         longitud        latitud           altitud      
 Min.   : 0.0000   Min.   :27.82   Min.   :-17.889   Min.   :   1.0  
 1st Qu.: 0.0000   1st Qu.:41.98   1st Qu.: -8.372   1st Qu.:  42.0  
 Median : 0.0000   Median :42.89   Median : -5.642   Median : 108.0  
 Mean   : 0.1246   Mean   :41.95   Mean   : -5.035   Mean   : 211.9  
 3rd Qu.: 0.0000   3rd Qu.:43.36   3rd Qu.: -2.906   3rd Qu.: 336.0  
 Max.   :81.0000   Max.   :43.57   Max.   :  4.216   Max.   :2400.0  

if (!empty_nodes) summary(df.cluster05)

   fecha_cnt          tmax            tmin            precip     
 Min.   :1.000   Min.   : -4.0   Min.   :-89.00   Min.   : 0.00  
 1st Qu.:2.000   1st Qu.:129.0   1st Qu.: 27.00   1st Qu.: 5.00  
 Median :3.000   Median :160.0   Median : 57.00   Median :12.00  
 Mean   :2.803   Mean   :156.4   Mean   : 54.03   Mean   :15.08  
 3rd Qu.:4.000   3rd Qu.:185.0   3rd Qu.: 82.00   3rd Qu.:22.00  
 Max.   :7.000   Max.   :281.0   Max.   :168.00   Max.   :80.00  
     nevada           prof_nieve          longitud        latitud       
 Min.   :0.00e+00   Min.   : 0.00000   Min.   :35.28   Min.   :-8.6494  
 1st Qu.:0.00e+00   1st Qu.: 0.00000   1st Qu.:38.99   1st Qu.:-5.2892  
 Median :0.00e+00   Median : 0.00000   Median :40.96   Median :-2.9553  
 Mean   :3.41e-05   Mean   : 0.09702   Mean   :40.51   Mean   :-2.5980  
 3rd Qu.:0.00e+00   3rd Qu.: 0.00000   3rd Qu.:42.08   3rd Qu.: 0.3664  
 Max.   :1.00e+00   Max.   :75.00000   Max.   :43.57   Max.   : 4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  58.0  
 Median : 353.0  
 Mean   : 401.2  
 3rd Qu.: 685.0  
 Max.   :1894.0  

if (!empty_nodes) summary(df.cluster06)

   fecha_cnt           tmax          tmin           precip           nevada 
 Min.   : 1.000   Min.   : 87   Min.   : 64.0   Min.   : 0.000   Min.   :0  
 1st Qu.: 3.000   1st Qu.:212   1st Qu.:150.0   1st Qu.: 0.000   1st Qu.:0  
 Median : 6.000   Median :233   Median :168.0   Median : 1.000   Median :0  
 Mean   : 6.486   Mean   :234   Mean   :169.4   Mean   : 6.219   Mean   :0  
 3rd Qu.: 9.000   3rd Qu.:259   3rd Qu.:196.0   3rd Qu.: 7.000   3rd Qu.:0  
 Max.   :12.000   Max.   :356   Max.   :244.0   Max.   :94.000   Max.   :0  
   prof_nieve           longitud        latitud          altitud     
 Min.   :0.0000000   Min.   :27.82   Min.   :-17.89   Min.   : 14.0  
 1st Qu.:0.0000000   1st Qu.:28.05   1st Qu.:-16.56   1st Qu.: 25.0  
 Median :0.0000000   Median :28.46   Median :-16.26   Median : 33.0  
 Mean   :0.0003521   Mean   :28.38   Mean   :-15.95   Mean   :130.9  
 3rd Qu.:0.0000000   3rd Qu.:28.48   3rd Qu.:-15.39   3rd Qu.: 64.0  
 Max.   :2.0000000   Max.   :28.95   Max.   :-13.60   Max.   :632.0  

if (!empty_nodes) summary(df.cluster07)

   fecha_cnt           tmax            tmin            precip           nevada 
 Min.   : 1.000   Min.   : 13.0   Min.   :-33.00   Min.   :  0.00   Min.   :0  
 1st Qu.: 4.000   1st Qu.: 94.0   1st Qu.: 23.00   1st Qu.:  0.00   1st Qu.:0  
 Median : 7.000   Median :129.0   Median : 51.00   Median :  3.00   Median :0  
 Mean   : 6.585   Mean   :138.4   Mean   : 60.83   Mean   : 10.29   Mean   :0  
 3rd Qu.:10.000   3rd Qu.:183.0   3rd Qu.: 98.00   3rd Qu.: 13.00   3rd Qu.:0  
 Max.   :12.000   Max.   :253.0   Max.   :159.00   Max.   :114.00   Max.   :0  
   prof_nieve         longitud        latitud         altitud    
 Min.   : 0.0000   Min.   :28.31   Min.   :-16.5   Min.   :2371  
 1st Qu.: 0.0000   1st Qu.:28.31   1st Qu.:-16.5   1st Qu.:2371  
 Median : 0.0000   Median :28.31   Median :-16.5   Median :2371  
 Mean   : 0.1766   Mean   :28.31   Mean   :-16.5   Mean   :2371  
 3rd Qu.: 0.0000   3rd Qu.:28.31   3rd Qu.:-16.5   3rd Qu.:2371  
 Max.   :46.0000   Max.   :28.31   Max.   :-16.5   Max.   :2371  

if (!empty_nodes) summary(df.cluster08)

   fecha_cnt         tmax             tmin            precip           nevada 
 Min.   : 1.0   Min.   :-12.00   Min.   :-72.00   Min.   : 16.00   Min.   :0  
 1st Qu.: 2.0   1st Qu.: 10.25   1st Qu.:-43.75   1st Qu.: 32.25   1st Qu.:0  
 Median : 2.5   Median : 24.00   Median :-29.50   Median : 50.00   Median :0  
 Mean   : 2.7   Mean   : 24.30   Mean   :-31.17   Mean   : 56.10   Mean   :0  
 3rd Qu.: 3.0   3rd Qu.: 35.75   3rd Qu.:-23.25   3rd Qu.: 71.50   3rd Qu.:0  
 Max.   :12.0   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve        longitud        latitud         altitud    
 Min.   : 357.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.: 481.5   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median : 599.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   : 685.7   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.: 765.2   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1], dim(df.cluster05)[1], dim(df.cluster06)[1], dim(df.cluster07)[1], dim(df.cluster08)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04", "cluster05", "cluster06", "cluster07", "cluster08"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.hist(df.cluster05)

if (!empty_nodes) mpr.hist(df.cluster06)

if (!empty_nodes) mpr.hist(df.cluster07)

if (!empty_nodes) mpr.hist(df.cluster08)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster05)

if (!empty_nodes) mpr.boxplot(df.cluster06)

if (!empty_nodes) mpr.boxplot(df.cluster07)

if (!empty_nodes) mpr.boxplot(df.cluster08)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
  df.cluster05.grouped <- mpr.group_by_geo(df.cluster05)
  df.cluster06.grouped <- mpr.group_by_geo(df.cluster06)
  df.cluster07.grouped <- mpr.group_by_geo(df.cluster07)
  df.cluster08.grouped <- mpr.group_by_geo(df.cluster08)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster05.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster06.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster07.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster08.grouped)

Visualización de 10 clústeres:

if (!empty_nodes) {
  plot(hac, hang=-1, labels=F)
  rect.hclust(hac, k=10)
}

Visualización de los clústers en el mapa

A qué clúster pertenece cada nodo del mapa de kohonen:

if (!empty_nodes) {
  groups <- cutree(hac, k=10)
  plot(model, type="mapping",
    bgcol=c("steelblue1","sienna1","yellowgreen","red","blue","yellow","purple","green","white","#1f77b4", '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2')[groups],
    shape = "straight", labels = "")
  add.cluster.boundaries(model, clustering=groups)
}

Análisis de las observaciones de cada cluster

if (!empty_nodes) {
  # Asignamos a cada registro su clúster
  df$cluster <- groups[model$unit.classif]
}

Nuevos dataframes por cluster

if (!empty_nodes) {
  # Creo nuevos dataframes, uno por cada clúster.
  df.cluster01 <- subset(df, cluster==1)
  df.cluster02 <- subset(df, cluster==2)
  df.cluster03 <- subset(df, cluster==3)
  df.cluster04 <- subset(df, cluster==4)
  df.cluster05 <- subset(df, cluster==5)
  df.cluster06 <- subset(df, cluster==6)
  df.cluster07 <- subset(df, cluster==7)
  df.cluster08 <- subset(df, cluster==8)
  df.cluster09 <- subset(df, cluster==9)
  df.cluster10 <- subset(df, cluster==10)

  # Extraigo del dataframe las features.
  df.cluster01 <- select(df.cluster01, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster02 <- select(df.cluster02, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster03 <- select(df.cluster03, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster04 <- select(df.cluster04, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster05 <- select(df.cluster05, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster06 <- select(df.cluster06, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster07 <- select(df.cluster07, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster08 <- select(df.cluster08, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster09 <- select(df.cluster09, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
  df.cluster10 <- select(df.cluster10, fecha_cnt, tmax, tmin, precip, nevada, prof_nieve, longitud, latitud, altitud)
}

if (!empty_nodes) summary(df.cluster01)

   fecha_cnt           tmax            tmin           precip          nevada 
 Min.   : 1.000   Min.   :126.0   Min.   : 13.0   Min.   : 0.00   Min.   :0  
 1st Qu.: 6.000   1st Qu.:232.0   1st Qu.:121.0   1st Qu.: 1.00   1st Qu.:0  
 Median : 8.000   Median :266.0   Median :148.0   Median : 7.00   Median :0  
 Mean   : 7.488   Mean   :264.9   Mean   :148.2   Mean   :10.69   Mean   :0  
 3rd Qu.: 9.000   3rd Qu.:297.0   3rd Qu.:176.0   3rd Qu.:16.00   3rd Qu.:0  
 Max.   :12.000   Max.   :403.0   Max.   :254.0   Max.   :95.00   Max.   :0  
   prof_nieve          longitud        latitud           altitud      
 Min.   : 0.00000   Min.   :35.28   Min.   :-8.6494   Min.   :   1.0  
 1st Qu.: 0.00000   1st Qu.:38.95   1st Qu.:-4.7667   1st Qu.:  43.0  
 Median : 0.00000   Median :40.84   Median :-2.1381   Median : 185.0  
 Mean   : 0.00322   Mean   :40.42   Mean   :-2.2238   Mean   : 344.4  
 3rd Qu.: 0.00000   3rd Qu.:41.98   3rd Qu.: 0.5706   3rd Qu.: 617.0  
 Max.   :35.00000   Max.   :43.57   Max.   : 4.2156   Max.   :1971.0  

if (!empty_nodes) summary(df.cluster02)

   fecha_cnt         tmax            tmin            precip     
 Min.   : 7.0   Min.   :  8.0   Min.   :-110.0   Min.   : 0.00  
 1st Qu.:11.0   1st Qu.:116.0   1st Qu.:  26.0   1st Qu.: 4.00  
 Median :11.0   Median :151.0   Median :  55.0   Median :11.00  
 Mean   :11.3   Mean   :151.2   Mean   :  58.6   Mean   :13.32  
 3rd Qu.:12.0   3rd Qu.:183.0   3rd Qu.:  90.0   3rd Qu.:20.00  
 Max.   :12.0   Max.   :305.0   Max.   : 189.0   Max.   :55.00  
     nevada            prof_nieve          longitud        latitud       
 Min.   :0.0000000   Min.   : 0.00000   Min.   :35.28   Min.   :-8.6494  
 1st Qu.:0.0000000   1st Qu.: 0.00000   1st Qu.:38.28   1st Qu.:-4.6800  
 Median :0.0000000   Median : 0.00000   Median :40.70   Median :-2.1381  
 Mean   :0.0001502   Mean   : 0.05624   Mean   :40.09   Mean   :-2.1095  
 3rd Qu.:0.0000000   3rd Qu.: 0.00000   3rd Qu.:41.66   3rd Qu.: 0.5706  
 Max.   :1.0000000   Max.   :59.00000   Max.   :43.57   Max.   : 4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  44.0  
 Median : 258.0  
 Mean   : 379.1  
 3rd Qu.: 667.0  
 Max.   :1894.0  

if (!empty_nodes) summary(df.cluster03)

   fecha_cnt           tmax             tmin            precip     
 Min.   : 1.000   Min.   : 69.00   Min.   :-15.00   Min.   : 3.00  
 1st Qu.: 2.000   1st Qu.: 76.00   1st Qu.: -5.00   1st Qu.: 9.00  
 Median : 2.000   Median : 91.00   Median : -1.00   Median :14.00  
 Mean   : 5.222   Mean   : 94.44   Mean   : 10.89   Mean   :12.89  
 3rd Qu.:12.000   3rd Qu.: 97.00   3rd Qu.: 30.00   3rd Qu.:15.00  
 Max.   :12.000   Max.   :140.00   Max.   : 35.00   Max.   :24.00  
     nevada        prof_nieve        longitud        latitud      
 Min.   :2.000   Min.   : 0.000   Min.   :40.48   Min.   :-3.450  
 1st Qu.:2.000   1st Qu.: 0.000   1st Qu.:40.48   1st Qu.:-3.450  
 Median :2.000   Median : 1.000   Median :41.67   Median :-1.033  
 Mean   :2.778   Mean   : 3.111   Mean   :41.14   Mean   :-2.107  
 3rd Qu.:3.000   3rd Qu.: 5.000   3rd Qu.:41.67   3rd Qu.:-1.033  
 Max.   :6.000   Max.   :13.000   Max.   :41.67   Max.   :-1.033  
    altitud     
 Min.   :263.0  
 1st Qu.:263.0  
 Median :263.0  
 Mean   :416.4  
 3rd Qu.:608.1  
 Max.   :608.1  

if (!empty_nodes) summary(df.cluster04)

   fecha_cnt           tmax             tmin             precip      
 Min.   : 1.000   Min.   :-53.00   Min.   :-121.00   Min.   :  0.00  
 1st Qu.: 3.000   1st Qu.: 27.00   1st Qu.: -35.00   1st Qu.: 20.00  
 Median : 6.000   Median : 66.00   Median :   1.00   Median : 32.00  
 Mean   : 6.425   Mean   : 72.75   Mean   :   5.09   Mean   : 36.21  
 3rd Qu.:10.000   3rd Qu.:117.00   3rd Qu.:  45.00   3rd Qu.: 48.00  
 Max.   :12.000   Max.   :218.00   Max.   : 128.00   Max.   :126.00  
     nevada    prof_nieve         longitud        latitud           altitud    
 Min.   :0   Min.   :  0.000   Min.   :40.78   Min.   :-4.0103   Min.   :1213  
 1st Qu.:0   1st Qu.:  0.000   1st Qu.:41.77   1st Qu.: 0.7292   1st Qu.:1894  
 Median :0   Median :  0.000   Median :42.47   Median : 0.9844   Median :2230  
 Mean   :0   Mean   :  7.832   Mean   :42.06   Mean   : 0.0693   Mean   :2160  
 3rd Qu.:0   3rd Qu.:  0.000   3rd Qu.:42.64   3rd Qu.: 1.5242   3rd Qu.:2400  
 Max.   :0   Max.   :343.000   Max.   :42.77   Max.   : 2.4378   Max.   :2535  

if (!empty_nodes) summary(df.cluster05)

   fecha_cnt           tmax            tmin            precip           nevada 
 Min.   : 1.000   Min.   : 42.0   Min.   :-42.00   Min.   : 20.00   Min.   :0  
 1st Qu.: 3.000   1st Qu.:127.0   1st Qu.: 58.00   1st Qu.: 47.00   1st Qu.:0  
 Median :10.000   Median :149.0   Median : 76.00   Median : 57.00   Median :0  
 Mean   : 7.743   Mean   :151.8   Mean   : 78.59   Mean   : 58.99   Mean   :0  
 3rd Qu.:11.000   3rd Qu.:174.0   3rd Qu.: 99.00   3rd Qu.: 69.00   3rd Qu.:0  
 Max.   :12.000   Max.   :269.0   Max.   :183.00   Max.   :110.00   Max.   :0  
   prof_nieve         longitud        latitud           altitud      
 Min.   : 0.0000   Min.   :28.48   Min.   :-16.329   Min.   :   1.0  
 1st Qu.: 0.0000   1st Qu.:42.24   1st Qu.: -7.860   1st Qu.:  42.0  
 Median : 0.0000   Median :43.12   Median : -5.642   Median : 108.0  
 Mean   : 0.0784   Mean   :42.06   Mean   : -4.959   Mean   : 198.5  
 3rd Qu.: 0.0000   3rd Qu.:43.37   3rd Qu.: -2.906   3rd Qu.: 336.0  
 Max.   :35.0000   Max.   :43.57   Max.   :  4.216   Max.   :1405.0  

if (!empty_nodes) summary(df.cluster06)

   fecha_cnt          tmax            tmin            precip     
 Min.   :1.000   Min.   : -4.0   Min.   :-89.00   Min.   : 0.00  
 1st Qu.:2.000   1st Qu.:129.0   1st Qu.: 27.00   1st Qu.: 5.00  
 Median :3.000   Median :160.0   Median : 57.00   Median :12.00  
 Mean   :2.803   Mean   :156.4   Mean   : 54.03   Mean   :15.08  
 3rd Qu.:4.000   3rd Qu.:185.0   3rd Qu.: 82.00   3rd Qu.:22.00  
 Max.   :7.000   Max.   :281.0   Max.   :168.00   Max.   :80.00  
     nevada           prof_nieve          longitud        latitud       
 Min.   :0.00e+00   Min.   : 0.00000   Min.   :35.28   Min.   :-8.6494  
 1st Qu.:0.00e+00   1st Qu.: 0.00000   1st Qu.:38.99   1st Qu.:-5.2892  
 Median :0.00e+00   Median : 0.00000   Median :40.96   Median :-2.9553  
 Mean   :3.41e-05   Mean   : 0.09702   Mean   :40.51   Mean   :-2.5980  
 3rd Qu.:0.00e+00   3rd Qu.: 0.00000   3rd Qu.:42.08   3rd Qu.: 0.3664  
 Max.   :1.00e+00   Max.   :75.00000   Max.   :43.57   Max.   : 4.2156  
    altitud      
 Min.   :   1.0  
 1st Qu.:  58.0  
 Median : 353.0  
 Mean   : 401.2  
 3rd Qu.: 685.0  
 Max.   :1894.0  

if (!empty_nodes) summary(df.cluster07)

   fecha_cnt          tmax            tmin            precip          nevada 
 Min.   : 1.00   Min.   : -8.0   Min.   :-58.00   Min.   : 90.0   Min.   :0  
 1st Qu.: 4.00   1st Qu.:125.0   1st Qu.: 63.00   1st Qu.:104.0   1st Qu.:0  
 Median :10.00   Median :148.5   Median : 83.50   Median :117.0   Median :0  
 Mean   : 8.35   Mean   :154.0   Mean   : 86.12   Mean   :126.7   Mean   :0  
 3rd Qu.:11.00   3rd Qu.:180.0   3rd Qu.:110.00   3rd Qu.:136.0   3rd Qu.:0  
 Max.   :12.00   Max.   :350.0   Max.   :223.00   Max.   :422.0   Max.   :0  
   prof_nieve         longitud        latitud           altitud       
 Min.   : 0.0000   Min.   :27.82   Min.   :-17.889   Min.   :   1.00  
 1st Qu.: 0.0000   1st Qu.:40.80   1st Qu.: -8.616   1st Qu.:  27.57  
 Median : 0.0000   Median :42.43   Median : -6.056   Median : 200.00  
 Mean   : 0.4316   Mean   :41.22   Mean   : -5.541   Mean   : 301.24  
 3rd Qu.: 0.0000   3rd Qu.:42.89   3rd Qu.: -2.039   3rd Qu.: 370.00  
 Max.   :81.0000   Max.   :43.57   Max.   :  3.035   Max.   :2400.00  

if (!empty_nodes) summary(df.cluster08)

   fecha_cnt           tmax          tmin           precip           nevada 
 Min.   : 1.000   Min.   : 87   Min.   : 64.0   Min.   : 0.000   Min.   :0  
 1st Qu.: 3.000   1st Qu.:212   1st Qu.:150.0   1st Qu.: 0.000   1st Qu.:0  
 Median : 6.000   Median :233   Median :168.0   Median : 1.000   Median :0  
 Mean   : 6.486   Mean   :234   Mean   :169.4   Mean   : 6.219   Mean   :0  
 3rd Qu.: 9.000   3rd Qu.:259   3rd Qu.:196.0   3rd Qu.: 7.000   3rd Qu.:0  
 Max.   :12.000   Max.   :356   Max.   :244.0   Max.   :94.000   Max.   :0  
   prof_nieve           longitud        latitud          altitud     
 Min.   :0.0000000   Min.   :27.82   Min.   :-17.89   Min.   : 14.0  
 1st Qu.:0.0000000   1st Qu.:28.05   1st Qu.:-16.56   1st Qu.: 25.0  
 Median :0.0000000   Median :28.46   Median :-16.26   Median : 33.0  
 Mean   :0.0003521   Mean   :28.38   Mean   :-15.95   Mean   :130.9  
 3rd Qu.:0.0000000   3rd Qu.:28.48   3rd Qu.:-15.39   3rd Qu.: 64.0  
 Max.   :2.0000000   Max.   :28.95   Max.   :-13.60   Max.   :632.0  

if (!empty_nodes) summary(df.cluster09)

   fecha_cnt           tmax            tmin            precip           nevada 
 Min.   : 1.000   Min.   : 13.0   Min.   :-33.00   Min.   :  0.00   Min.   :0  
 1st Qu.: 4.000   1st Qu.: 94.0   1st Qu.: 23.00   1st Qu.:  0.00   1st Qu.:0  
 Median : 7.000   Median :129.0   Median : 51.00   Median :  3.00   Median :0  
 Mean   : 6.585   Mean   :138.4   Mean   : 60.83   Mean   : 10.29   Mean   :0  
 3rd Qu.:10.000   3rd Qu.:183.0   3rd Qu.: 98.00   3rd Qu.: 13.00   3rd Qu.:0  
 Max.   :12.000   Max.   :253.0   Max.   :159.00   Max.   :114.00   Max.   :0  
   prof_nieve         longitud        latitud         altitud    
 Min.   : 0.0000   Min.   :28.31   Min.   :-16.5   Min.   :2371  
 1st Qu.: 0.0000   1st Qu.:28.31   1st Qu.:-16.5   1st Qu.:2371  
 Median : 0.0000   Median :28.31   Median :-16.5   Median :2371  
 Mean   : 0.1766   Mean   :28.31   Mean   :-16.5   Mean   :2371  
 3rd Qu.: 0.0000   3rd Qu.:28.31   3rd Qu.:-16.5   3rd Qu.:2371  
 Max.   :46.0000   Max.   :28.31   Max.   :-16.5   Max.   :2371  

if (!empty_nodes) summary(df.cluster10)

   fecha_cnt         tmax             tmin            precip           nevada 
 Min.   : 1.0   Min.   :-12.00   Min.   :-72.00   Min.   : 16.00   Min.   :0  
 1st Qu.: 2.0   1st Qu.: 10.25   1st Qu.:-43.75   1st Qu.: 32.25   1st Qu.:0  
 Median : 2.5   Median : 24.00   Median :-29.50   Median : 50.00   Median :0  
 Mean   : 2.7   Mean   : 24.30   Mean   :-31.17   Mean   : 56.10   Mean   :0  
 3rd Qu.: 3.0   3rd Qu.: 35.75   3rd Qu.:-23.25   3rd Qu.: 71.50   3rd Qu.:0  
 Max.   :12.0   Max.   : 76.00   Max.   :  8.00   Max.   :180.00   Max.   :0  
   prof_nieve        longitud        latitud         altitud    
 Min.   : 357.0   Min.   :40.78   Min.   :-4.01   Min.   :1894  
 1st Qu.: 481.5   1st Qu.:40.78   1st Qu.:-4.01   1st Qu.:1894  
 Median : 599.0   Median :40.78   Median :-4.01   Median :1894  
 Mean   : 685.7   Mean   :40.78   Mean   :-4.01   Mean   :1894  
 3rd Qu.: 765.2   3rd Qu.:40.78   3rd Qu.:-4.01   3rd Qu.:1894  
 Max.   :1834.0   Max.   :40.78   Max.   :-4.01   Max.   :1894  

Número de elementos en cada clúster

if (!empty_nodes) {
  df.clusters.dim <- c(dim(df.cluster01)[1], dim(df.cluster02)[1], dim(df.cluster03)[1], dim(df.cluster04)[1], dim(df.cluster05)[1], dim(df.cluster06)[1], dim(df.cluster07)[1], dim(df.cluster08)[1], dim(df.cluster09)[1], dim(df.cluster10)[1])
  barplot(df.clusters.dim,
          names.arg = c("cluster01", "cluster02", "cluster03", "cluster04", "cluster05", "cluster06", "cluster07", "cluster08", "cluster09", "cluster10"),
          col = "steelblue1")
}

Distribución de los datos

if (!empty_nodes) mpr.hist(df.cluster01)

if (!empty_nodes) mpr.hist(df.cluster02)

if (!empty_nodes) mpr.hist(df.cluster03)

if (!empty_nodes) mpr.hist(df.cluster04)

if (!empty_nodes) mpr.hist(df.cluster05)

if (!empty_nodes) mpr.hist(df.cluster06)

if (!empty_nodes) mpr.hist(df.cluster07)

if (!empty_nodes) mpr.hist(df.cluster08)

if (!empty_nodes) mpr.hist(df.cluster09)

if (!empty_nodes) mpr.hist(df.cluster10)

if (!empty_nodes) mpr.boxplot(df.cluster01)
if (!empty_nodes) mpr.boxplot(df.cluster02)

if (!empty_nodes) mpr.boxplot(df.cluster03)

if (!empty_nodes) mpr.boxplot(df.cluster04)

if (!empty_nodes) mpr.boxplot(df.cluster05)

if (!empty_nodes) mpr.boxplot(df.cluster06)

if (!empty_nodes) mpr.boxplot(df.cluster07)

if (!empty_nodes) mpr.boxplot(df.cluster08)

if (!empty_nodes) mpr.boxplot(df.cluster09)

if (!empty_nodes) mpr.boxplot(df.cluster10)

Localización geográfica de las estaciones de medida de cada cluster

# Agrupa por longitud y latitud para rellenar el mapa con menos datos.
if (!empty_nodes) {
  df.cluster01.grouped <- mpr.group_by_geo(df.cluster01)
  df.cluster02.grouped <- mpr.group_by_geo(df.cluster02)
  df.cluster03.grouped <- mpr.group_by_geo(df.cluster03)
  df.cluster04.grouped <- mpr.group_by_geo(df.cluster04)
  df.cluster05.grouped <- mpr.group_by_geo(df.cluster05)
  df.cluster06.grouped <- mpr.group_by_geo(df.cluster06)
  df.cluster07.grouped <- mpr.group_by_geo(df.cluster07)
  df.cluster08.grouped <- mpr.group_by_geo(df.cluster08)
  df.cluster09.grouped <- mpr.group_by_geo(df.cluster09)
  df.cluster10.grouped <- mpr.group_by_geo(df.cluster10)
}

if (!empty_nodes) mpr.draw_map(spain, df.cluster01.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster02.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster03.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster04.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster05.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster06.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster07.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster08.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster09.grouped)

if (!empty_nodes) mpr.draw_map(spain, df.cluster10.grouped)